Rescaling the Generated Output to the pixel intensity range in DCGAN

I have a question about the DCGAN tutorial

I was wondering why the Generator does not rescale the generated output to the range [0,1] or [0, 255] (the range of pixel intensity) for each channel before sending it to the Discriminator during the training process.

I thought that if the Generator model did not rescale the generated output, the Discriminator could easily learn the input images that did not include the value in the range [0,1] or [0, 255] as the fake image.